Usage Considerations
The API_Request in-database function is used for model scoring and inference calculations.
Before using the API_Request in-database function, the AWS endpoint of Amazon SageMaker or analytic model should have been trained and deployed on AWS.
You also need the endpoint address, region and AWS credentials (which have permissions to use this in-database function) to execute a query to score Vantage data with this AWS analytic service.
Usage Example: Use SageMaker Endpoint for Scoring Vantage Data
In this example, an Amazon SageMaker model has been deployed to the AWS US-East-2 region with the endpoint 'sagemaker-xgboost-2021-10-20-15-43-44-623'. The authentication fields are intentionally left as ‘replace with your AWS credentials’. The expected input fields to this AWS analytic model are in the Vantage table ‘NEW_FINANCIAL_TRANS’.
SELECT rec_id, output as fraud_risk_score FROM tapidb.API_Request ( ON NEW_FINANCIAL_TRANS USING AUTHORIZATION('{"Access_ID":"replace with your AWS Access ID", "Session_Token":"replace with your AWS Session Token", "Region":"us-east-2"}') ENDPOINT('sagemaker-xgboost-2021-10-20-15-43-44-623') CONTENT_TYPE('csv') KEY_START_INDEX('1') ) as DT ;
API_Request query result:
rec_id fraud_risk_score 597417 0.8734374642372131 392307 0.5901673436164856 52268 0.9538228511810303 76809 0.47017282247543335 744410 0.9500066637992859 751493 0.5542758703231812 676641 0.881127119064331 146036 0.9368776082992554 631585 0.42199018597602844 45837 0.4777362048625946 51034 0.2817979156970978 484301 0.9431955814361572 603917 0.9265019297599792 554445 0.7330561280250549 12892 0.6860849857330322